Professional Documents
Culture Documents
I. INTRODUCTION
I
I diff =
j =1
(1)
I
I rest =
j =1
I rest =
[4]
(2)
[2]
(3)
[3]
(4)
n
n
j =1
( )
I rest = max I j
(5)
R (t ) = 1 F (t )
(6)
1) System Equivalent
The system equivalent seen from the busbar at which the
transformer protection is analyzed was modeled with a
balanced three-phase voltage source behind an RL coupled
element represented in symmetrical components. The values of
the positive and zero sequence resistance and reactance used in
this study (see Table I) were obtained using the short-circuit
analysis software ANAFAS [8], which contains a complete
and accurate description of entire Brazilian power system.
TABLE I SYSTEM EQUIVALENT
R(pu)
X(pu)
Positive Sequence
2.6288
9.3253
Zero Sequence
6.5806
20.648
R (t ) = e t
(8)
IV. DEVELOPMENTS
In this study, the differential protection of a power
transformer is tested using the Monte Carlo method and
2) Power transformer
The power transformer whose differential protection is
evaluated in this study is used at the high voltage distribution
system of COPEL, which is the major power utility company
operating in the state of Paran, south of Brazil. It is a threephase, 41 MVA, 138/34.5/13.8 kV power transformer,
supplied from the 138-kV side, which feeds distribution
networks of 34.5 kV and 13.8 kV. Since its winding
configuration is Yg-yg-, zig-zag grounding transformers are
installed at the low voltage busbar to give a reference to
ground protective relays installed at this side.
The power transformer was modeled following the same
principles used in the hybrid model available in ATP [9].
However, the following simplifications and/or improvements
were performed:
TABLE II
GROUND FAULT
ATP (Simplified [A])
ATP (full [A])
Phase ANAFAS
Currents (A) Currents (A) Error (%) Currents (A) Error (%)
A
679.1
420.6
38.07
678.6
0.07
B
679.1
420.6
38.07
678.5
0.09
C
679.1
421.5
37.93
679.0
0.01
V(kV)
I(A)
Energization>1_?-IAW1(A)
Energization>2_?-IBW 1(A)
15.57
7.46
Energization>3_?-ICW 1(A)
2
Magnitude (Mag)
-2
a +b i + c i
i
1
i1
1
= i2
M
1 M
n in1
11
1
2
= a i + b + c i
1
i1 2
a
1
i2 2 b
~
y = X
M
M c
1
1 in 2
1
~
~
~
~
= [( X T X ) 1 X T ] y = P y
(9)
-4
120
140
160
180
Time (ms)
200
Electrotek Concepts
(10)
220
240
(a)
400
[A]
300
200
(11)
100
-100
-200
-300
-400
0,00
(file energ3.pl4; x-var t) c:X0002A-PRIA
0,02
c:X0002B-PRIB
0,04
0,06
0,08
[s]
0,10
c:X0002C-PRIC
(b)
resistance and
specifications.
reactance were
100
100
I diff 2 +
I diff 4
PCT 2
PCT 4
(12)
where Idiff2 e Idiff4 are the second and the fourth harmonic
contents of the differential current. PCT2 and PCT4 are the
Value
60
160
400
2.91
4.36
4.36
11
11
0
0.35
30
60
6
10
20
10
Y
Y
a N (0, )
a = 3 = N (0, )
a
4 max U (0,1)
a
5 max U (0,1)
a
6 max U (0,1)
(13)
=p
2 =
p (1 p )
n
(14)
(15)
p (1 p )
(16)
TABLE VIII
SUCCESS RATE OBTAINED USING DIFFERENT HARMONIC REFERENCE SETTINGS
IN THE HARMONIC RESTRAINT TECHNIQUE (SEE DEFINITION OF PCT2 AND
PCT4 IN TABLE VI)
Number of
PCT2
PCT4
Success Rate (%)
Simulations
20
15
1176
86.39 3.00
20
10
847
90.67 3.00
15
10
327
96.64 3.00
B. Turn-to-ground faults
Table IX shows the success rate obtained for turn-to-ground
faults using two different minimum differential settings in a
percentage restraint differential protection scheme. In the
analysis, the same set of parameters listed in Table VI was
considered, except for parameter O87P, which varied from
0.35 to 0.5. According to Table IX, a reduction in the value of
the minimum differential current increases the reliability of the
differential function during turn-to-ground faults. However, it
increases the risk of misoperation due to noise in the
secondary circuit of current transformers. Furthermore, the
obtained values suggest that percentage restraint differential
protection is possibly not a good technique to detect turn-toground faults.
model seems to be sufficiently accurate, although improvements could be made provided additional information on the
relay algorithm were made available by its manufacturer.
VII. CONCLUSIONS
This paper presents a time-domain methodology to evaluate
the reliability of a transformer protection differential function.
To account for the random nature of many factors that affect
the protection behavior, concepts related to the Monte Carlo
method and reliability theory were used. The proposed method
seems promising in the sense that it gives to the user a more
precise idea of the expected behavior of the differential relay
under different conditions. Moreover, it allows the user to test
many different relay settings to choose the ones that are more
adequate for a given application. Future studies with more
detailed transformer and relay models, as well as with the
inclusion of current transformer models, are in due course.
VIII. REFERENCES
[1]
TABLE IX
SUCCESS RATE OBTAINED USING DIFFERENT HARMONIC REFERENCE
SETTINGS IN A PERCENTAGE RESTRAINT DIFFERENTIAL PROTECTION SCHEME
(SEE DEFINITION OF PARAMETER O87P IN TABLE VI)
O87P
Number of simulations
Success Rate (%)
0.35
1831
75.87 3.00
0.5
2116
69.61 3.00
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]